Hybridization is important in evolution, but is speciation?

Abstract

Few contributions have influenced the field of evolutionary biology more than Ernst Mayr’s authoritative book entitled Systematics and The Origin of Species (Mayr, 1942). Here, the author went further than his forerunners in depicting the living world as consisting of discrete clusters of reproductively isolated entities called species and evolutionary differentiation as occurring essentially through a series of speciation events in geographical isolation. Gene flow and hybridization were seen as mainly destructive forces with little evolutionary consequence except in counteracting or retarding evolutionary divergence. Ernst Mayr’s ideas imply that speciation, the process in which populations become reproductively isolated from each other, is the key event in evolutionary diversification. Only when gene flow is brought to a halt are organisms free to diverge unhindered. In the current issue of Journal of Evolutionary Biology, Abbott et al. (2013) tell a very different story. During the ongoing genetic and genomic revolution, students of speciation have come to learn that Mayr’s view is at best incomplete. Hybridization between genetically differentiated populations is widespread, common and has evolutionary consequences. In many circumstances, hybridization plays an integral part in the diversification process and may significantly enhance differentiation. Abbott et al. (2013) present an impressive series of arguments and references to recent empirical and theoretical results on the importance of hybridization in evolution. I wonder, however, whether Mayr’s intellectual universe still affects our interpretations of results and our way of thinking. Can we any longer take for granted that speciation, the final cut-off of gene flow between differentiated taxa, really is this key event in evolution? Could it be that by focusing on speciation as the goal of population divergence or at least as a most significant milestone, we fail to appreciate the full array of processes that result directly or indirectly from hybridization and the consequences these have on diversification and biological organization? My intention here is not to debate species concepts or to question whether species are real. Biodiversity is certainly not a blur of endless forms. Phenotypes are organized into recognizable, relatively discrete clusters, and the evidence is overwhelming that reproductive barriers are important in structuring biodiversity (see e.g. Coyne & Orr, 2004). My intention is rather to focus on the process of evolutionary diversification with an open mind. A great deal of diversification occurs prior to the final completion of reproductive isolation, and as Abbott et al. (2013) point out, some 10–30% of our animal and plant species are known to hybridize and exchange genes with others on a regular basis. A further large proportion of our extant nonhybridizing species (being geographically or reproductively isolated from their closest living relatives at present) are likely to have hybridized some time in the past, even repeatedly. Hence, understanding the evolutionary effects of all this hybridization and gene exchange is utterly important. Also, it is important to investigate the consequences speciation (the final sealing of gene exchange) actually has on the diversification process and in structuring biodiversity. One possible outcome of hybridization is that the introgression of selectively favoured alleles from one population into the other can bring together new adaptive combination of alleles that increase diversification. Abbott et al. (2013) make a five-point argument for why hybridization may enhance adaptive differentiation. In short, they argue that hybridization may act as a possibly more abundant source of adaptive genetic variation than mutation because mutations are rare and hybridization common. They cite Grant & Grant (1994) who estimated that the amount of new, additive genetic variance introduced by hybridization was two to three orders of magnitude higher than that introduced by mutation in Darwin’s finches. We may add one more difference between a mutated allele and one introduced by hybridization. The mutated allele has been altered randomly, whereas the one introduced by hybridization has been shaped by natural selection, albeit in a differentiated genome (deleterious mutations have been purged and any beneficial mutations gone to fixation by selection). Intuitively, I would therefore think that an allele introduced by hybridization on average is more likely to do something good for the organism it enters than a mutated one. The relative distribution of fitness effects of mutated versus introgressed alleles is an interesting empirical question that should be investigated further. Moreover, not only single alleles, but co-adapted suites of interacting genes could potentially be transferred through introgressive hybridization, leading to rapid origin of complex adaptations and evolutionary novelties. Introgression as a source of new genetic variation is by definition not available for completely reproductively isolated species. Hence, if in the long run hybridization and Correspondence: Glenn-Peter Saetre, Department of Biology, Centre for Ecological and Evolutionary Synthesis (CEES), PO Box 1066 Blindern, N-0316 Oslo, Norway. Tel.: +47 22857291; fax: +47 22854001; e-mail: g.p.satre@bio.uio.no